The objective of this program is to develop improved non-invasive clinical methods for the analysis of pulmonary sounds. These methods will be based on postulated models of the physical mechanisms responsible for sound production. Correlation between clinically-observed sounds and the physical models should yield improved specificity regarding the type and extent of pulmonary disease. Rales ("crackles"), rhonchi ("Wheezes"), and other adventitious respiratory sounds depend upon physical lung properties such as airway elasticity, the presence of fluid in the airways, ventilation rate, lung inflation, patient posture, and the sound transmission characteristics of the lung and chest wall. Predictive models relating these physical properties to the lung sounds produced by specific disease states do not exist. Our program seeks to develop such physical models and validate them using clinical data obtained from patients with confirmed diagnoses. In the first year, we have postulated two models for crackle production that will be tested and refined by further work. We have observed crackles in a normal person using a special filtering and amplification system to process the microphone signal. We are developing a high speed, computer-based multichannel sound analyzer to simultaneously study sounds from multiple chest sites.